> You seem to presume:  1)  the costs of a PWD-offered solution
> are predominately attributable
> to the cost of the selected mainframe virtual machine, and 2)
> there is no reason for IBM's
> management to prefer one solution over another, and 3) that
> the solutions are exchangable.

Actually, if you're willing to share the breakdown in your costs between
HW, emulation and IBM software, I'd be very interested in seeing it.
With regard to your other two points, I'll agree that we hold differing
opinions -- the work we have done with Hercules has proved it sufficient
for our needs, which at the end of the day, is what most people decide
on.  If this were a perfect world, perhaps things might be different,
but for my needs, Hercules is quite sufficient.

> The irony of your reply is we are talking about a development
> environment for a commercial
> version of sendmail, a product which, using the analysis
> you've supplied here, would have
> little to say for itself even though we know it has a valued
> place in the Linux enterprise
> firmament.   As far as I can tell, IBM has yet to hear from
> even one developer of commercial
> software who would otherwise meet the conditions of software
> loan that the present PWD
> offering is prohibitive.

Here's where I think we're missing each other. You are talking
specifically about commercial development. Sure, Sendmail, Inc probably
has enough cash to afford a medium-sized FlexES based solution. I don't
doubt that it would be a good choice for them.  I'm talking about the
one-man shops and the shops that need an occasional test system or a
small development system that are NOT in the business of producing
mass-market software.  *Those* people can reasonably afford one or two
fairly high-powered PCs, but not necessarily a dedicated system just for
emulated 390 HW, and their budget for tools and toys does not extend to
chunks of cash > $10K.  While this may not seem to be a desirable
market, I'd argue that this is the marketplace that has driven the
survival of Windows -- and of Linux Intel -- and if the 390 is to
continue to play a large role or attract new developers and new
applications, then these people are the ones you need to court and
attract, not the big vendors.

Consider also the success of the LCDS -- IBM has other offerings for
ISVs to do porting, etc, and when Joe Dempster and I started campaigning
for a open development system, all sorts of hissy fits were thrown about
lost revenue, but eventually we prevailed on IBM to make the system open
to everyone without charge. It's the small shops that have driven the
LCDS to succeed. Why? Because they can get access to a interesting
technology like a genuine 390 w/o a huge investment (from their
perspective).  Did IBM make money on that? No... but that's exactly what
we're discussing here -- a method to gain some additional revenue
without a lot of additional expense, and that benefits IBM by a) showing
that they're serious about attracting new developers and new blood to
the 390 platform, and b) forestalling a piracy problem in the making by
making it easy to *be* legal and aboveboard (avoiding the RIAA mistake
with MP3 files: letting technology get so far ahead of them that they
can't put the genie back in the bottle). Most people are basically
honest and will be reasonable about licensing if their wallets can
support it.

IBM has to make the ADCDs anyway for the PID folks, and I don't see
anything wrong with giving the PID folks nice hardware discounts and
other perks that CD-only licensees don't get. I'm trying to make a case
that IBM should offer the ADCD for licensing and use elsewhere at a
reasonable price for a small developer.  That was the point of my
comment earlier -- if I could license the ADCD for a price in the
ballpark of buying the MS development tools (plus or minus 10-15% or
so), I'd buy a legal copy for each of my developers, and both IBM and I
walk away from the deal happy (and I'd bet that the ADCD production
process will get somewhat cheaper due to economies of scale). I'm
getting tools I need and want at a price I can afford, and IBM is
getting good revenue and a proactive stance on controlling their rights
to their software without having to do negative stuff like go sue
people.

> I can assure you that were you able to convince IBM of a need
> for a developer platform and
> deployment model addressing a developer class beyond those
> who qualify for its program
> today, then Flex-ES could be packaged to meet that need
> within the bounds of costs you
> yourself have declared reasonable for such developers to pay.

OK, I'll put it out in the open. Can you provide a software only FlexES
package including the ADCD license in the $2.5-3K range? If so, then we
should talk seriously about it. If not, then see above for the argument
that there are people that do not fit the current program and cannot
afford to participate.

-- db

Reply via email to